Serious fault has been found[133] with Bergson’s attempt to establish, by scientific research in the subject of aphasia, the ontological independence of spirit, the seat of memory, from body. But on other grounds than such scientific investigation the issue of this attempt appears to me at best a futile achievement; for the result is in any case the reinstatement, untouched, of that problem of all radical dualism, a problem which Bergson solves only by metaphor whose brilliance may be luminous itself, but has no illumination for the problem, which is how reactive states are also conscious. There is a theory which relates consciousness and matter to each each other as the opposite sides of a surface in relief. The objection to this “double aspect” theory that has weighed most, in criticism, is that the ground of the parallelism between convexity and concavity--to wit, a logical implication of each other--is obviously absent in the parallelism of consciousness and matter. Whatever parallelism experience actually finds between them is not deducible from either concept: there is nothing in the definition of the sensation blue to suggest an afferent nervous current; nothing in the latter to suggest a sensation. They are incommensurate. But when you conceive convexity, in that fact you conceive concavity also, and _vice versa_. They are related as plus and minus. The objection appeals to analysis of the definition of consciousness or of matter, or challenges the advocate of the theory to study his sensation or his neural process and see if there be in either of them anything of the other. A difficulty which immediately arises when this challenge is accepted has been understood to be decisive against the theory. It is this: Any definition of consciousness which the advocate of the theory may propose as the concept to be analyzed must, in order to fulfil the first requirement of logical definition, be in terms of that which is not consciousness. And this seems to the critic to beg the question. If you define consciousness so, he objects, you make its definition imply matter; but there is then nothing of consciousness in it; what you have got is only matter. That is to assume an equation between them. You state the value of _x_ in terms of _y_, but then you haven’t got _x_, but only _y_. It is otherwise with terms that really have the correlation you claim for consciousness and matter. Thus you can equate convexity with concavity in terms of either alone, as _m_ = -(-_m_). In this there is no assumption. But what you say of _x_ is that it equals _ay_, which is something _distinguishable_ from _x_ and whose equality to _x_ is just the problem.